Update app.py
Browse files
app.py
CHANGED
@@ -344,7 +344,8 @@ def Compute_ALL_Model_memory(vocab_size, layer_num, hidden_size, ffn_size, seq_l
|
|
344 |
numParameters = round(Get_BillionParameter(numParameters), 3)
|
345 |
model_states_memory = round(Get_GigaByte(model_states_memory), 3)
|
346 |
activation_memory = round(Get_GigaByte(activation_memory), 3)
|
347 |
-
|
|
|
348 |
|
349 |
# record
|
350 |
new_row = pd.DataFrame([[layer_num, hidden_size, ffn_size, seq_length, head_num, group_query_num, dp, tp, pp, cp, gpu_num, b, is_fp8,
|
@@ -363,6 +364,7 @@ def Compute_ALL_Model_memory(vocab_size, layer_num, hidden_size, ffn_size, seq_l
|
|
363 |
Model parameters on each device = {str(numParameters)} B, \n
|
364 |
Model_states = Weight + Gradient + Optimizer = {str(model_states_memory)} GB, \n
|
365 |
Activation = {str(activation_memory)} GB, \n
|
|
|
366 |
Total memory consumption = {str(Total)} GB \n
|
367 |
""", record_df, count
|
368 |
|
|
|
344 |
numParameters = round(Get_BillionParameter(numParameters), 3)
|
345 |
model_states_memory = round(Get_GigaByte(model_states_memory), 3)
|
346 |
activation_memory = round(Get_GigaByte(activation_memory), 3)
|
347 |
+
other_memory = 5
|
348 |
+
Total = round(model_states_memory + activation_memory + other_memory, 3)
|
349 |
|
350 |
# record
|
351 |
new_row = pd.DataFrame([[layer_num, hidden_size, ffn_size, seq_length, head_num, group_query_num, dp, tp, pp, cp, gpu_num, b, is_fp8,
|
|
|
364 |
Model parameters on each device = {str(numParameters)} B, \n
|
365 |
Model_states = Weight + Gradient + Optimizer = {str(model_states_memory)} GB, \n
|
366 |
Activation = {str(activation_memory)} GB, \n
|
367 |
+
Other memory = 5 GB, \n
|
368 |
Total memory consumption = {str(Total)} GB \n
|
369 |
""", record_df, count
|
370 |
|